Late life depression (LLD) is an increasingly important cause of disability and mortality worldwide. Clinical studies have characterized many aspects of the phenomenology, and treatment response in LLD, but the etiology of LLD remains unclear. Few models have yielded testable hypotheses, but an emerging concept is that inflammation may have particular relevance, especially in late life depression. We have preliminary data supporting increased inflammatory cytokines in LLD compared with controls that are associated with decreased hippocampus and amygdala volumes. LLD also had specific impairment in neuropsychological function in episodic memory and executive function and decreased hippocampus and amygdala resting state functional connectivity. Antidepressant treatment improved IL-6 levels, neuropsychological testing and resting state connectivity. In the current proposal we propose to recruit patients with LLD (n = 100) and controls (n = 50) matched for vascular risk factors to characterize immune function, brain structure and connectivity and neuropsychological function. Further, we will randomize participants with LLD to 10 weeks of SSRI treatment (n=50) or no treatment (wait list) (n=50), and assess participants pre- and post-treatment with peripheral and central cytokine levels, neuropsychological tests and brain resting state functional connectivity. Aim 1: Characterize neuroanatomical and neuropsychological correlates of abnormal cytokines in LLD. Hypothesis 1: a) Compared with matched controls, LLD will have abnormalities in peripheral and CSF inflammatory cytokines, neuropsychological function (e.g., executive function, episodic memory), hippocampal, amygdala and prefrontal cortex (PFC) structure and functional connectivity; b) Peripheral/central inflammatory cytokine levels will be associated with volume loss in hippocampus, including CA2-3, PFC, and amygdala; and c) Cumulative duration of depression will correlate with volume loss in hippocampus, including CA2-3, amygdala, and PFC. Aim 2: Characterize the reversibility of brain dysfunction, including cognitive impairment, with treatment of late- life depression and its association with inflammation. Hypothesis 2: a) Subjects randomized to treatment will have greater normalization of IL-6 and IL-10, greater improvements in memory and executive function and improvements in resting state functional connectivity compared with those randomized to initial no treatment; b) Improvement in clinical depressive scores will correlate with normalizatio of inflammatory cytokines. Significance: Demonstration of hypothesized links between neurobiology, depression and inflammation should augment development of treatment strategies for this debilitating disorder.